Blunted apoptosis of erythrocytes in mice deficient in the heterotrimeric G-protein subunit Gαi2

Putative functions of the heterotrimeric G-protein subunit Gαi2-dependent signaling include ion channel regulation, cell differentiation, proliferation and apoptosis. Erythrocytes may, similar to apoptosis of nucleated cells, undergo eryptosis, characterized by cell shrinkage and cell membrane scrambling with phosphatidylserine (PS) exposure. Eryptosis may be triggered by increased cytosolic Ca2+ activity and ceramide. In the present study, we show that Gαi2 is expressed in both murine and human erythrocytes and further examined the survival of erythrocytes drawn from Gαi2-deficient mice (Gαi2−/−) and corresponding wild-type mice (Gαi2+/+). Our data show that plasma erythropoietin levels, erythrocyte maturation markers, erythrocyte counts, hematocrit and hemoglobin concentration were similar in Gαi2−/− and Gαi2+/+ mice but the mean corpuscular volume was significantly larger in Gαi2−/− mice. Spontaneous PS exposure of circulating Gαi2−/− erythrocytes was significantly lower than that of circulating Gαi2+/+ erythrocytes. PS exposure was significantly lower in Gαi2−/− than in Gαi2+/+ erythrocytes following ex vivo exposure to hyperosmotic shock, bacterial sphingomyelinase or C6 ceramide. Erythrocyte Gαi2 deficiency further attenuated hyperosmotic shock-induced increase of cytosolic Ca2+ activity and cell shrinkage. Moreover, Gαi2−/− erythrocytes were more resistant to osmosensitive hemolysis as compared to Gαi2+/+ erythrocytes. In conclusion, Gαi2 deficiency in erythrocytes confers partial protection against suicidal cell death.

Similar to nucleated cells, erythrocytes may undergo suicidal death or eryptosis 19,20 , which, similar to apoptosis, is triggered by osmotic shock and characterized by cell shrinkage and cell membrane scrambling 20,21 . Eryptosis may be triggered by activation of Ca 2+ -permeable cation channels 20 which subsequently leads to increase of cytosolic Ca 2+ . The molecular identity of these cation channels has not been completely characterized but apparently involves TRPC6 channels 22 . The cation channels are activated by prostaglandin E 2 , which is formed following exposure of erythrocytes to hyperosmotic shock 19 . The channels are further activated by a wide variety of cell stressors, xenobiotics and endogenous mediators 19 . Ca 2+ activates Ca 2+ -sensitive K + channels with exit of KCl and osmotically obliged water and thus cell shrinkage 19,20 . An increase of cytosolic Ca 2+ is further followed by stimulation of cell membrane scrambling with exposure of phosphatidylserine (PS) at the cell surface 19,20 . The Ca 2+ sensitivity of cell membrane scrambling is further enhanced by ceramide 21 . PS-exposing cells are bound to macrophages, engulfed and degraded and thus cleared from circulating blood 19,20,[23][24][25] . To the best of our knowledge, the impact of Gα i2 on erythrocyte survival and suicidal death has hitherto not been reported.
In the present study we explored whether the Gα i2 isoform is expressed in erythrocytes and whether it participates in the regulation of erythrocyte survival. To this end, eryptosis was determined in erythrocytes from Gα i2 knockout mice (Gαi2 −/− ) and their wild type littermates (Gαi2 +/+ ).
Immunoblotting was employed to examine whether Gα i2 is expressed in human and murine erythrocytes. To this end, erythrocytes from humans or from mice were isolated and purified. Equal amounts of protein lysates were immunoblotted. GAPDH served as a loading control. As depicted in Fig. 2A, the incubation with Gα i2 specific antibodies yielded a single band of 40 kDa in human erythrocytes as well as erythrocytes from Gαi2 +/+ mice, but not in erythrocytes from Gαi2 −/− mice. The bands appearing below 40 kDa are presumably the result of non-specific antibody binding. Densitometry analysis revealed that Gα i2 protein is significantly more abundant in mouse erythrocytes as compared to human erythrocytes (Fig. 2B). Thus, Gα i2 is expressed in both human and murine erythrocytes.
Next, we explored whether Gα i2 deficiency influences erythrocyte survival. To this end, using annexin V binding, forward scatter and Fluo3 fluorescence in FACS analysis we analyzed erythrocyte cell membrane PS exposure, cell shrinkage and cytosolic Ca 2+ activity, respectively. As depicted in Fig. 3A, freshly drawn and untreated erythrocytes were visualized using confocal microscopy and quantification of multiple fields showed a decreased ratio of annexin V binding cells to total cells (observed under transmission light) per field in Gαi2 −/− erythrocytes (0.028 ± 0.007; n = 4) as compared to Gαi2 +/+ erythrocytes (0.069 ± 0.007; n = 4). PS exposure was simultaneously quantified using FACS analysis (50,000 cells were quantified) and confirmed that in both freshly drawn blood (Fig. 3B,C) and following 12 h incubation in Ringer solution (Fig. 3D), the percentage of annexin V binding erythrocytes was significantly lower in Gαi2 −/− mice than in Gαi2 +/+ mice. Quantification of forward scatter showed that the cell volume was significantly larger in Gαi2 −/− erythrocytes as compared to Gαi2 +/+ erythrocytes (Fig. 4A,B). Both cell membrane PS exposure and cell shrinkage are influenced by cytosolic Ca 2+ activity 20 . As shown in Fig. 4C,D, the percentage of Fluo3 positive erythrocytes was slightly but significantly lower in Gαi2 −/− mice as compared to Gαi2 +/+ mice. These data suggest an inhibitory effect of Gα i2 deficiency on eryptosis.
Further experiments then addressed the susceptibility of Gα i2-deficient erythrocytes to osmotic shock ex vivo, a pathophysiological cell stressor and a known stimulator of eryptosis. As illustrated in Fig. 5A,B, exposure of erythrocytes for 30 min to hyperosmotic Ringer (550 mM sucrose was added to the Ringer solution to reach the final osmolarity of 850 mOsm), significantly enhanced PS exposure, an effect, however, significantly blunted in Gαi2 −/− erythrocytes as compared to Gαi2 +/+ erythrocytes. Erythrocyte forward scatter was quantified to determine hyperosmotic shock-triggered cell shrinkage. As shown in Fig. 5C,D, forward scatter was significantly reduced by hyperosmotic shock in erythrocytes from both Gαi2 −/− and Gαi2 +/+ mice. The effect was significantly less pronounced in Gαi2 −/− erythrocytes than in Gαi2 +/+ erythrocytes.
To elucidate the mechanism contributing to the protective effect of Gα i2 deficiency against hyperosmotic shock-triggered eryptosis, we determined erythrocyte cytosolic Ca 2+ activity following hyperosmotic shock. As shown in Fig. 6A,B, exposure of erythrocytes to hyperosmotic shock significantly enhanced the percentage of Fluo3 positive erythrocytes. The effect was, however, significantly blunted in Gαi2 −/− erythrocytes as compared to Gαi2 +/+ erythrocytes. Further experiments explored the resistance of erythrocytes to a decline of extracellular osmolarity. As illustrated in Fig. 6C, the resistance of erythrocytes to graded decrease of osmolarity was significantly lower in Gαi2 +/+ than in Gαi2 −/− erythrocytes. Thus, Gα i2 deficiency counteracts the sensitivity of erythrocytes to both hyper-and hypoosmotic shock.

Discussion
The present observations disclose the expression of Gα i2 in human and murine erythrocytes and further reveals that Gα i2 deficiency confers partial protection against suicidal erythrocyte death or eryptosis. Our findings show that the percentage of eryptotic cells in circulating blood is significantly lower in Gαi2 −/− mice as in Gαi2 +/+ mice. Gαi2 −/− mice do not show overt changes in erythrocyte parameters such as erythrocyte count, hematocrit, hemoglobin concentration and reticulocyte count. The impact of Gα i2 deficiency on erythrocytes is unmasked in the presence of pathophysiological cell stressors ex vivo such as hyperosmotic shock and following treatment with C6 ceramide and bacterial sphingomyelinase, whereby eryptosis is significantly less pronounced in Gαi2 −/− erythrocytes as compared to Gαi2 +/+ erythrocytes.
Our data show that in the absence of stress, the difference between the percentage of PS-exposing erythrocytes in Gαi2 +/+ mice and Gαi2 −/− mice is subtle (~0.2%) yet statistically significant. Previous studies have shown that spontaneous PS exposure in freshly drawn erythrocytes from healthy wild-type mice of different strains does not exceed 1% 19 of the total number of circulating erythrocytes. Thus, in transgenic mice which display a phenotype of reduced eryptosis, the percentage of PS-exposing circulating erythrocytes may be significantly lower than in wild-type mice despite relatively lower magnitudes of difference. Exposure of erythrocytes to hypertonic extracellular environment in vitro simulates the osmotic conditions encountered in the kidney medulla 20 . In conditions such as acute renal failure, erythrocytes may enter eryptosis due to their entrapment in the kidney medulla 21 . Gα i2 deficiency may blunt eryptosis and thus favorably influence the respective clinical condition. Our data show that, in addition to curtailing PS exposure, Gαi2 −/− erythrocytes showed increased resistance to cell shrinkage following hyperosmotic shock. Accordingly, the mean corpuscular cell volume was significantly larger in Gαi2 −/− erythrocytes. Along those lines, it is intriguing to speculate that Gα i2 influences cell volume regulatory ion channels in erythrocytes.
Mechanistically, hyperosmotic shock is a powerful stimulator of Ca 2+ entry and ceramide formation in erythrocytes 20 . We observed that following hyperosmotic shock of erythrocytes, Gα i2 deficiency leads to subtle but significant decrease of cytosolic Ca 2+ entry. On the other hand, Gα i2 may additionally mediate hyperosmotic shock-induced eryptosis by influencing ceramide signaling 21 . This is corroborated by our data showing a mitigating effect of Gα i2 deficiency on eryptosis triggered by either C6 ceramide or bacterial sphingomyelinase. Ceramide sensitizes erythrocytes to the eryptotic effect of enhanced Ca 2+ concentration and may stimulate eryptosis without appreciable increase in cytosolic Ca 2+ activity 27 . Ceramide further modifies the interaction of the erythrocyte membrane with the cytoskeleton thereby increasing membrane fragility 28 . As Gα i2 is an essential regulator for Ca 2+ signaling in nucleated cells, it is possible that the inhibitory effect of Gα i2 deficiency on erythrocyte death is, at least in part, mediated by its influence on cytosolic Ca 2+ activity.
Eryptosis is inhibited by catecholamines including dopamine 29 . Interestingly, dopamine-dependent signaling involves pertussis toxin-sensitive Gα i2 30 . Further signaling molecules that regulate the eryptosis machinery include AMPK 20 , p38 MAPK 31 , CK1α 32 , PAK2 33 , PDK1 20 , MSK1/2 34 and CDK4 35 . Eryptosis is triggered by a myriad of xenobiotics and endogenous substances 20,36-48 , and accelerated eryptosis contributes to the anemia associated with several clinical disorders 20 , including iron deficiency 49 , sepsis 50 , renal failure 51 , hepatic failure 52 , malignancy 24 , ageing 53 and Wilson's disease 54 . Eryptotic erythrocytes adhere to the vascular wall 55 , and stimulate blood clotting 56 . Excessive eryptosis may thus interfere with microcirculation and participate in the vascular injury of metabolic syndrome 57 . Accordingly, Gαi2 −/− mice may be particularly resistant to derangements of microcirculation following exposure to triggers of eryptosis. Moreover, eryptosis has been shown to influence the quality of stored erythrocytes 58 . Pharmacologically targeting Gα i2, at least in theory, may further provide new avenues in the treatment of conditions associated with anemia resulting from increased eryptosis 20 . On the other hand, Gα i2 modulation may serve as a novel target for the treatment of malaria, a condition where eryptosis plays a protective role in ameliorating parasitemia by expediting the clearance of pathogen-infected erythrocytes 20 . In conclusion, the G-protein subunit Gα i2 is expressed in human and murine erythrocytes and participates in the regulation of erythrocyte suicide.

Materials and Methods
Mice. Experiments were performed in Gα i2 knockout mice (Gαi2 −/− ) and their wild type littermates (Gαi2 +/+ ) of 6-9 weeks of age. The mice were generated and initially characterized on a SV129 background 18 . Mice were backcrossed on a C57BL6 background and kept under specified pathogen-free (SPF) environment in individually ventilated cages (IVC) to prolong life expectancy 4,59 . All animal experiments were conducted according to the German law for the care and use of laboratory animals and were approved by local government authorities (Regierungspräsidium Tübingen).

Blood count, incubation and solutions.
For all experiments except for the blood count, heparin blood was retrieved from the retrobulbar plexus of mice. For the blood count, EDTA blood was analyzed using an electronic hematology particle counter (type MDM 905 from Medical Diagnostics Marx; Butzbach, Germany) equipped with a photometric unit for haemoglobin determination. Plasma erythropoietin levels were determined using an immunoassay kit according to the manufacturer's instructions (R&D Systems, Wiesbaden, Germany). Murine erythrocytes were isolated by being washed two times with Ringer solution containing (in mM): 125 NaCl, 5 KCl, 1 MgSO 4 , and 32 HEPES/NaOH (pH 7.4), 5 glucose, and 1 CaCl 2 . Where indicated, sucrose (550 mM), C6 ceramide (50 μ M; Sigma) or bacterial sphingomyelinase (0.01 U/ml; Sigma) were added to the Ringer solution. May-Grünwald staining was used to examine changes in erythrocyte shape. Briefly, 20 μ l of erythrocytes were smeared and fixed using methanol onto a glass slide, and stained with 5% Giemsa Azur-Eosin (Merck Millipore, Germany) in phosphate-buffered saline (in mM: 1.05 KH 2 PO 4 , 2.97 Na 2 HPO 4 , 155.2 NaCl) for 20 min. Subsequently, images were taken on a Nikon Diaphot 300 Microscope (Nikon Instruments, Germany).

Reticulocyte count and markers of erythrocyte maturation.
For determination of the reticulocyte count EDTA-whole blood (2.5 μ l) was added to 500 μ l Retic-COUNT (Thiazole orange) reagent from Becton Dickinson. Samples were stained for 30 min at room temperature, and flow cytometry was performed according to the manufacturer's instructions. Forward scatter (FSC), side scatter (SSC), and thiazole orange-fluorescence intensity Determination of the osmotic resistance. Two microliters of blood were added to 200 μ l of PBS solutions with decreasing osmolarity. After centrifugation for 5 min at 3000 rpm, the supernatant was transferred to a Immunoblotting. To examine the expression of Gα i2 in human or murine erythrocytes, 150 μ l erythrocyte pellet was lysed in 50 ml of 20 mM HEPES/NaOH (pH 7.4). Ghost membranes were pelleted (15,000 g for 20 min at 4 °C) and lysed in 200 μ l lysis buffer (50 mM Tris-HCl, pH 7.5; 150 mM NaCl; 1% Triton X-100; 0.5% SDS; 1 mM NaF; 1 mM Na 3 VO 4 ; and 0.4% β -mercaptoethanol) containing protease inhibitor cocktail (Sigma, Schnelldorf, Germany). Triton X-100, a non-ionic detergent, was used in erythrocyte ghost preparation due to its effective solubilization power and a relatively mild effect on membrane-bound enzymes 60 . In all cases, 60 μ g of protein was solubilized in Laemmli sample buffer at 95 °C for 5 min and resolved by pre-casted 10% SDS-PAGE gel (Invitrogen, Karlsruhe, Germany). For immunoblotting, proteins were electrotransferred onto a polyvinylidene difluoride (PVDF) membrane and blocked with 5% nonfat milk in TBS-0.10% Tween 20 at room temperature for 1 h. Then, the membrane was incubated with anti-G-protein alpha inhibitor 2 antibody (1:5000; 40 kDa; Abcam Cat# ab157204) at 4 °C overnight. After being washed (in TBS-0.10% Tween 20) and subsequently blocked, the blots were incubated with secondary anti-rabbit antibody (1:2000; Cell Signaling) for 1 h at room temperature. After being washed, the antibody binding was detected with the ECL detection reagent (Amersham, Freiburg, Germany).
Confocal microscopy and immunofluorescence. For the visualization of eryptotic erythrocytes, 4 μ l of erythrocytes, incubated in the respective experimental solutions, were stained with FITC-conjugated annexin-V (1:100 dilution; ImmunoTools, Friesoythe, Germany) in 200 μ l Ringer solution containing 5 mM CaCl 2 . Then, the erythrocytes were washed twice and finally resuspended in 50 μ l of Ringer solution containing 5 mM CaCl 2 . Twenty μ l were mounted with Prolong Gold antifade reagent (Invitrogen, Darmstadt, Germany) onto a glass slide and covered with a coverslip. Sections were analyzed using a Leica TCS-SP / Leica DM RB confocal laser scanning microscope. Images were processed with Leica Confocal Software LCS (Version 2.61).
Statistics. Data are expressed as arithmetic means ± SEM, and statistical analysis was made using ANOVA or t-test, as appropriate. n denotes the number of different erythrocyte specimens studied.